schema-to-data v0.1.0

schema-to-data

Straightforward algorithm for generating mock data given a json-schema. schema-to-data exists to serve schema-to-generator which provides a much more flexible api for generating mock data. If you are just here to generate mock data then install schema-to-generator instead!

Note: This library has not been released to npm yet and is still under development

Installation

npm install --save-dev https://github.com/randograms/schema-to-data

Getting Started

const {
  schemaToData,
  createWithDefaults, // see "Working with Custom Defaults" below
} = require('schema-to-data');

// prints a string
console.log(schemaToData({ type: 'string' }));

// prints an object with at least 7 properties
console.log(schemaToData({
  type: 'object',
  minProperties: 7,
}));

// prints either an integer or decimal that is greater than or equal to 70 OR a string array with at least 2 items
console.log(schemaToData({
  type: ['number', 'array'],
  minimum: 70,
  items: { type: 'string' },
  minItems: 2,
}));

Features

Github project roadmap.

Currently Supported Keywords

• All
  • allOf
  • anyOf
  • const
  • enum
  • oneOf
  • type
• Array
  • additionalItems
  • items
  • maxItems
  • minItems
• Object
  • additionalProperties
  • maxProperties
  • minProperties
  • patternProperties
  • properties
  • propertyNames
  • required
• String
  • format
    • date
    • date-time
    • email
    • ipv4
    • ipv6
    • time
    • uuid
  • maxLength
  • minLength
  • pattern
• Number
  • maximum
  • minimum

Planned Supported Keywords

• All
  • default
  • examples
• Array
  • contains
• String
  • format
    • idn-email
    • hostname
    • idn-hostname
    • uri
    • uri-reference
    • iri
    • iri-reference
    • uri-template
    • json-pointer
    • relative-json-pointer
    • regex
• Number
  • exclusiveMaximum (draft4 and draft6)
  • exclusiveMinimum (draft4 and draft6)
  • multipleOf

Unplanned Keywords

• All
  • not
• Array
  • uniqueItems
• Object
  • dependencies

Custom Defaults

Creating a Custom schemaToData Instance

const { createWithDefaults } = require('schema-to-data');

const customSchemaToData = createWithDefaults({
  minStringLength: 50,
});

// prints a string with at least 50 characters
console.log(customSchemaToData({ type: 'string' }));

Inspecting Instance Defaults

const {
  schemaToData,
  createWithDefaults,
} = require('schema-to-data');

const customSchemaToData = createWithDefaults({
  stringLengthRange: 1000,
});

// prints default defaults
console.log(schemaToData.getDefaults());

// prints instance defaults
console.log(customSchemaToData.getDefaults());

Available createWithDefaults Options

Schema keywords always take precedence over configurable defaults (ex: if a schema has "minLength", then "minLength" will be used instead of "minStringLength").

Project Scope

Goals

• Provide a clear algorithm for generating mock data given a json-schema
  • Maintain separation of concerns via thorough unit tests
  • Provide documentation for the internals of the library at both a high level and a more granular level
  • Maintain high level integration tests
  • Maintain integration regression tests for found issues
• Provide clear expectations based on json-schema draft versions
  • Limit extraneous default functionality that is not stated in the json-schema specification
  • Be open to configurable quality of life enhancements
  • Separate supported behavior based on json-schema draft

Out of Scope Concerns

• Schema validation
  • schema-to-data should ignore any malformed json-schema keywords and will ignore any custom json-schema keywords
  • schema-to-data will always attempt to generate valid data until it is forced to throw an exception
  • Consider using Another JSON Schema Validator for schema validation as it is well-documented and thoroughly versed in the nuances of json-schema
  • schema-to-data officially uses Another JSON Schema Validator for its tests
• Extreme edge cases
  • Validating data against a json-schema is much easier than generating data, therefore edge cases will only be handled if a solution can fit cleanly into the algorithm
  • Consider using schema-to-generator instead of schema-to-data as it provides more flexibility when working with complex schemas
• Seeded "random" data
  • schema-to-data does not provide a means to seed random data even if some of the underlying libraries provide such functionality
• Exposing internal dependencies
  • schema-to-data is not strictly tied to any underlying library for data generation and will not expose any underlying library
• Parsing $ref
  • There are other libraries that can parse $ref before the schema is passed to schema-to-data

More Docs

Developer Guide